TAILIEUCHUNG - Báo cáo khoa học: "Fast Consensus Decoding over Translation Forests"

The minimum Bayes risk (MBR) decoding objective improves BLEU scores for machine translation output relative to the standard Viterbi objective of maximizing model score. However, MBR targeting BLEU is prohibitively slow to optimize over k-best lists for large k. In this paper, we introduce and analyze an alternative to MBR that is equally effective at improving performance, yet is asymptotically faster — running 80 times faster than MBR in experiments with 1000-best lists. | Fast Consensus Decoding over Translation Forests John DeNero Computer Science Division University of California Berkeley denero@ David Chiang and Kevin Knight Information Sciences Institute University of Southern California chiang knight @ Abstract The minimum Bayes risk MBR decoding objective improves BLEU scores for machine translation output relative to the standard Viterbi objective of maximizing model score. However MBR targeting BLEU is prohibitively slow to optimize over k-best lists for large k. In this paper we introduce and analyze an alternative to MBR that is equally effective at improving performance yet is asymptotically faster running 80 times faster than MBR in experiments with 1000-best lists. Furthermore our fast decoding procedure can select output sentences based on distributions over entire forests of translations in addition to k-best lists. We evaluate our procedure on translation forests from two large-scale state-of-the-art hierarchical machine translation systems. Our forest-based decoding objective consistently outperforms k-best list MBR giving improvements of up to BLEU. 1 Introduction In statistical machine translation output translations are evaluated by their similarity to human reference translations where similarity is most often measured by BLEU Papineni et al. 2002 . A decoding objective specifies how to derive final translations from a system s underlying statistical model. The Bayes optimal decoding objective is to minimize risk based on the similarity measure used for evaluation. The corresponding minimum Bayes risk MBR procedure maximizes the expected similarity score of a system s translations relative to the model s distribution over possible translations Kumar and Byrne 2004 . Unfortunately with a non-linear similarity measure like BLEU we must resort to approximating the expected loss using a k-best list which accounts for only a tiny fraction of a model s full posterior distribution. In this .

TAILIEUCHUNG - Chia sẻ tài liệu không giới hạn
Địa chỉ : 444 Hoang Hoa Tham, Hanoi, Viet Nam
Website : tailieuchung.com
Email : tailieuchung20@gmail.com
Tailieuchung.com là thư viện tài liệu trực tuyến, nơi chia sẽ trao đổi hàng triệu tài liệu như luận văn đồ án, sách, giáo trình, đề thi.
Chúng tôi không chịu trách nhiệm liên quan đến các vấn đề bản quyền nội dung tài liệu được thành viên tự nguyện đăng tải lên, nếu phát hiện thấy tài liệu xấu hoặc tài liệu có bản quyền xin hãy email cho chúng tôi.
Đã phát hiện trình chặn quảng cáo AdBlock
Trang web này phụ thuộc vào doanh thu từ số lần hiển thị quảng cáo để tồn tại. Vui lòng tắt trình chặn quảng cáo của bạn hoặc tạm dừng tính năng chặn quảng cáo cho trang web này.